Back to Case Studies

October 2026 · 9 min read

Case Study

Chasing the Winner: $10,000 Into Each Year's Top S&P 500 Stock Since 2000

What if you bought last year's best-performing S&P 500 stock, every year since 2000? The result is spectacular, and the catch is the best part.

Cover: $10,000 into each year's best stock since 2000, $2.9 million? A log-scale chart of the strategy against the S&P 500.

Hey everyone,

Welcome to the first LTLM Case Study. The idea is simple: take a question investors love to ask, run it against real market history, and show every number. This one is a classic. What if you just bought last year's best-performing S&P 500 stock, every single year? No stock-picking skill, no research. Just buy the winner, hold it for a year, then sell and move into the next winner. I ran it from the end of 2000 through October 2026, and the result is both spectacular and a little bit of a trap. Let's get into it.

1. The experiment

Here are the rules I used:

  • The starting point. $10,000 invested on the last trading day of 2000.
  • The pick. On the last trading day of each year, find the stock with the highest price return that calendar year and put everything into it. That's 2000's winner for 2001, 2001's winner for 2002, and so on.
  • The rotation. Hold for exactly one calendar year, sell, and roll all of it into the new winner. 26 years of this, with 2026 counted year to date through October 12.
  • The universe. Companies in the S&P 500 today, ranked by calendar-year price return. A stock also has to have traded for the entire year it won, so I skipped stocks that listed mid-year, such as Leidos in 2006 and SanDisk in 2025. I also skipped the 2018 listing for TKO, whose history only exists under that ticker since 2023, and used the next stock, DexCom.
  • The scorecard. Price returns only: no dividends, trading costs or taxes. I compare it to the S&P 500's price return (using the SPY fund as the proxy).

2. The result

On paper, it's astonishing. The winner-rotation strategy turned $10,000 into about $2.92M, a compound return of roughly 24.4% a year. The S&P 500, over the same stretch, turned it into about $59,257, or roughly 7.1% a year, before dividends.

Rotate into the top stock

$2.92M

Per year, compounded

+24.4%

S&P 500 (price only)

$59,257

Years it beat the index

18 of 26

Growth of $10,000, 2000 to 2026 (log scale)

$1K $10K $100K $1M $10M 2000 2005 2010 2015 2020 2025 $2.92M $59,257 2011: $1,731 2004 peak: $13,868
Each year's top S&P 500 stockS&P 500 (price only)

The chart uses a log scale, where each gridline is ten times the one below it, because otherwise the first twenty years would be a flat line next to the last five. And that flat line is exactly where the story gets interesting.

3. Year by year

Here is every pick. The third column is the huge gain that made it the winner. The fourth is what it did once we owned it. Notice how rarely the two look alike.

Year heldStock boughtIts gain the year beforeIts gain while heldYour $10,000 becomesS&P 500
2001LHLaboratory Corp. of America+377%−7.6%$9,240$8,710
2002BKNGBooking Holdings (then Priceline)+324%−72.6%$2,532$6,724
2003ELVElevance Health (then Anthem)+153%+19.6%$3,028$8,479
2004AXONAxon (then TASER)+1,934%+358.0%$13,868$9,208
2005AXONAxon (then TASER)+358%−78.5%$2,982$9,485
2006MNSTMonster Beverage (then Hansen Natural)+322%+67.3%$4,988$10,784
2007ICEIntercontinental Exchange+196%+77.6%$8,859$11,129
2008FSLRFirst Solar+786%−48.8%$4,536$6,867
2009DLTRDollar Tree+61%+15.3%$5,230$8,480
2010AMDAdvanced Micro Devices+342%−16.4%$4,372$9,566
2011NFLXNetflix+216%−60.4%$1,731$9,547
2012DPZDomino's Pizza+111%+36.1%$2,356$10,835
2013REGNRegeneron+203%+56.0%$3,676$14,054
2014TSLATesla+330%+48.5%$5,459$15,642
2015SWKSSkyworks Solutions+159%+5.4%$5,754$15,516
2016NFLXNetflix+133%+13.6%$6,536$17,006
2017AMDAdvanced Micro Devices+309%−10.0%$5,883$20,305
2018XYZBlock (then Square)+151%+61.6%$9,506$19,026
2019DXCMDexCom+108%+86.9%$17,767$24,505
2020CVNACarvana+191%+157.0%$45,662$28,475
2021TSLATesla+731%+46.9%$67,077$36,164
2022DVNDevon Energy+192%+50.6%$101,018$29,112
2023CEGConstellation Energy+129%+37.6%$139,001$36,186
2024CVNACarvana+1,001%+294.0%$547,663$44,617
2025APPAppLovin+722%+103.5%$1.11M$51,934
2026*LITELumentum+338%+162.4%$2.92M$59,257

* 2026 is year to date through October 12. Stock returns are price returns from StatMuse; S&P 500 returns from ChartRow (SPY price). Dollar values assume the full balance is moved each year. Not every year's winner is a household name today: companies were renamed, merged or reorganized, so I show the name they traded under then where it differs.

4. What the numbers actually say

  • The winning year was never the repeat. The average winner gained about 380% in the year it topped the list. The median gain in the year we owned it was about 42%. Last year's rocket is usually this year's coin flip.
  • The first 17 years lost money. After 2017, our $10,000 was worth $5,883, down about 41%, while the S&P 500 had turned it into $20,305. Nearly all the profit came from 2018 on.
  • It took a brutal ride. Seven of the 26 years were losers: LH in 2001, Priceline in 2002, TASER in 2005, First Solar in 2008, AMD in 2010 and 2017, and Netflix in 2011. Three of them lost more than 60%. From the end-2004 peak of $13,868, the account fell to $1,731 by 2011, a drop of about 87.5%. It didn't set a new high until 2019.
  • The last three years did the heavy lifting. About 96% of the total gain came in 2024, 2025 and 2026, when Carvana, AppLovin and Lumentum went up 294%, 104% and 162%. Take away those three and this is a very average story.
  • Repeat winners happen. Axon (when it was TASER), Netflix, AMD, Tesla and Carvana each made the list twice. That's momentum showing up, and it helped, but it also didn't prevent TASER's 78% loss in 2005.

5. Why you couldn't actually run this

The headline number is real arithmetic on real prices, but it is not a strategy anyone could have followed. Four big reasons:

  • Survivorship and hindsight. I ranked stocks that are in the S&P 500 today. In 2003, TASER was a tiny company nowhere near the index, and Carvana in 2019 was a recent IPO still losing money. The list quietly contains companies that were later rewarded for winning, which makes the past look more repeatable than it was.
  • Taxes and costs. Selling every year in a taxable account turns each gain into a tax bill, and short-term gains are taxed at ordinary income rates. A 24% pre-tax return could look quite different after tax, and the larger the account, the more trading in thinly traded names costs.
  • Size and liquidity. Moving into a small-cap rocket works with $10,000. At $2 million it's a different problem.
  • Human nature. Stomaching an 87.5% drawdown, then handing everything to whatever just went up the most, is something almost no investor sticks with. Most would have quit after 2002 or after 2011, right before the best years.

6. What I'd take from it

  • Last year's winner is a headline, not a plan. Big one-year gains tell you a stock got hot, not that it will keep going.
  • Concentration cuts both ways. One stock at a time gave the best and worst outcomes here. The index never had a year worse than about −38%.
  • Time and sequence matter. The same returns in a different order could have meant a very different result, and 17 years of waiting is a long time to be underwater.
  • The boring option held up better than it looks. A plain S&P 500 fund grew $10,000 into about $59,257 with none of this drama. If you want more of the basics, start with the Investing Basics course and the handbook guide on S&P 500 holding periods.

The bottom line: this is a fun experiment about the power of compounding and the danger of hindsight. What you do with your own money is a question only you can answer.

NOT FINANCIAL ADVICE. This post is an educational case study of historical data and is not a recommendation to buy, sell, or hold any security. Past performance does not predict future results. Returns are price-only and hypothetical, assume no taxes, fees or slippage, and draw on stocks that are S&P 500 members today. Figures come from third-party data providers, are rounded, and may differ slightly from other sources; check before relying on them. This post was prepared with AI assistance; see our AI Disclosure.

Sources